Research On Control Method Of Quadruped Robot In Complex Terrain

Compared to wheeled and tracked robots,quadruped robots have superior terrain passing capabilities,making them promising in various fields such as expedition search and rescue,industrial inspection,and military reconnaissance.However,to achieve motion control of quadruped robots,several challenges in perception positioning,motion planning,and underlying control need to be addressed.Currently,existing quadruped robot control technology can only passively perceive environmental changes through feedback data from body joint encoders and IMU,making it difficult to actively adapt to the changing environment,resulting in poor passability on complex terrains.Additionally,quadruped robots are nonlinear,multiple-input multiple-output,and underactuated,posing a challenging task in achieving robust control.To address these challenges,thesis proposes a hierarchical control framework combining perception,planning,and control for quadruped robots in complex terrains.The framework,based on the ROS system,comprises four parts: terrain perception,state estimation,local planning,and underlying motion control.The terrain perception component generates a 2.5D grid map around the quadruped robot in real-time,providing prior information for local planning.The state estimation provides the system feedback.Local planning is mainly composed of two parts: the gait planner and nonlinear model predictive control.The gait planner is responsible for scheduling the running gait,selecting the optimal foothold point in complex terrain,and planning the swing trajectory of the foot.The nonlinear model predictive control optimizes the trajectory of the robot’s center of mass and solves the mechanical input sequence of the foot.The underlying motion control relies on the whole-body hybrid control theory to achieve the multi-task priority control of the quadruped robot.The experiments validated that the proposed control method can enable the quadruped robot to reach 4.15,driving it through rough terrain such as non-paved roads,discontinuous terrain such as plum piles,and step terrain such as stairs.The control method effectively ensures the motion stability of the quadruped robot under complex terrain conditions,improving its adaptability to complex environments,and providing excellent technical support for its industrial field applications.